AI: the future that is already here

Hon. Dr. Dory Reiling is a retired senior judge of the Amsterdam District Court with extensive experience in judicial reform and digital innovation. She has advised national and international bodies on artificial intelligence in the justice sector and contributed to key ethical frameworks, including the Council of Europe’s Ethical Charter on the use of AI in judicial systems. 

______________________________

In her earlier post, Dr. Keri Grieman concluded that artificial intelligence (AI) and the judiciary have an interesting future ahead of them, with some distinct choices to be made. In this post, I will look into some issues and choices that are already here but have been underexposed until now. The most pressing issues are independence and governance.

In 2024, a judge in in a European jurisdiction mentioned in their judgment how they had used ChatGPT to determine the remaining lifespan of solar panels. The tech community was elated - the courts are using AI! The legal community, however, was not so amused, as the case raised several issues. More about that later.

As a judge, I am always looking for facts and evidence. So, let us start with some facts about the use of AI in courts around the world.

AI in courts – a very diverse list

In Europe, in 2023, some countries reported to the Consultative Council of European Judges that some of them had a government strategy, some judiciaries had a vision, a plan for measures, some initiatives, even pilot projects, ideas or chatbots, and one judiciary was setting up a thinktank.

In one instance, the Ministry of Justice (MoJ), managing technology for the courts, decided to introduce an AI tool into the courts system. The idea was to have it used for summarizing large case files. The judiciary was not consulted, and some judges objected. So, the MoJ withdrew the tool, again without consulting the judiciary. Elsewhere in Europe, the local Bar commissioned a system to handle bankruptcy proceedings since neither the judiciary nor the MoJ had done so. The source code and the data are owned by the third party who built the system; individual judges are given access only to their own cases, and the third party now offers the data for sale on the market. In another jurisdiction, profiling judges is now a crime after the government decided the judiciary needed to open up its data.

Across the Atlantic, a North American court management group listed how they were experimenting: chat boxes to relieve court staff from answering routine questions from self-representing litigants; several sentencing algorithms providing robust guidance and information for the judges;  the use of translation tools for customer service (not in the courtroom); and speech to text applications used at a low level due to difficulties in recognizing accents. So far, no broad application of AI tools.

By far, the most interesting developments are happening in the courts in Asia, so let us have a look at some of the forerunners there. One judiciary experiments with generative AI (large language models) to help self-representing litigants with court documentation. Another has started an initiative with live transcription, with its Chief Justice advocating for ethical AI integration in legal research. A third uses sentencing prediction incidentally, despite ethical concerns raised by lawyers. In another country, the judiciary is actively integrating AI tools to enhance legal processes and support lawyers’ access to courts.

One jurisdiction has launched a Smart Courts Reform program under a government AI strategy. The courts use AI for recording testimonies, analyzing case materials, verifying database information and AI-powered automation. Smart courts, analyzing case materials and verifying database information, monitor and register every action. A smart system can flag a sensitive case, and thus ensure oversight by the court leaders.

Issues

These examples illustrate that there are issues AI presents for judiciaries and courts, and that those issues are already with us.

One issue is how to deal with AI-generated information, be it case information, search results or evidence. Fair procedure requires that the parties in a case be given an opportunity to present arguments about factual data not in the case file but found on the internet of the judge’s own accord. This is already standard case law in most countries.

Some particularly pressing issues are: How much room is there for judges to decide equitably? How can courts and judges be in control of their processes and accountable for their judgments? And the most important question: Who is in charge? To my mind, this is the most important question for courts and judiciaries implementing information technology in general, and artificial intelligence in particular. If individual judges need to be independent to decide impartially, how does AI affect that independence, and the institutional independence of courts and judiciaries? If liberal democracies want to safeguard their judiciaries, they need to be explicit about the impartiality and independence they want to deliver, and adapt the constitutional set-up accordingly.

Some of these examples show issues that have been around for a long time, and introducing AI will only compound the issues around impartiality and independence. So, let us think about how technology, including AI, impacts those values, encoded in human rights conventions, and that judiciaries are expected to deliver: impartiality, independence, public justice, access, and fair proceedings within a reasonable time.

Values

For a long time, implementing technology was framed in terms of New Public Management:  increasing efficiency, cost reduction and timeliness as the main values. But does that make for good justice? Judiciaries, when deciding about technology, AI or more traditional, need to be guided by the duty to deliver good justice: impartial, independent, public, accessible, fair justice, within a reasonable time. They also need to use technology to realise and increase the delivery of those justice values, not just to reduce cost for the government and the taxpayer. By the way: there is enough evidence that introducing technology does not reduce cost.

I will concentrate on impartiality and independence, its precondition, as they are the most urgent. Important aspects of independence and impartiality are governance (who decides what?), accountability (how to publicly explain decisions and judgments?) and access to information (who gets to see what?). Technology is no longer a mere tool; it has become an environment that supports delivering justice. Judiciaries are morphing into IT organizations that deliver justice. Forms of AI can become part of this environment. The general consensus by now is that AI should remain under human control, as advocated by the 2018 Ethical Charter on the use of artificial intelligence (AI) in judicial systems and their environment, developed by the European Commission for the Efficiency of Justice (CEPEJ).This human control can ensure respect for those values. Judiciaries and courts should institutionalize this control, and that demands new forms of governance.

To safeguard individual judicial impartiality and independence, judiciaries need to set up a framework for dealing with information about individual judges’ performance, profiles or otherwise. Who has access, and for what can they be used?

To maintain institutional judicial impartiality and independence in digital environments, judiciaries, whether through Councils of the Judiciary or otherwise, ought to be in charge, or at least effectively involved in the purchase, design and control of technology. They ought also to agree and participate in its introduction and implementation. This means they need adequate, sustainable resources for training and developing guidance. This issue is particularly important where the responsibility for court management rests with Ministries of Justice or is a matter of partnership between the judiciary and the Ministry of Justice. As the Consultative Council of European Judges observed in its Opinion 26: “Provision should also be made for judges to be kept up to date with technological innovation to facilitate their effective involvement and, where necessary, concurrence in the use of new and evolving technology.”

New reorientation on the values courts and judiciaries deliver, and their governance in particular, are just a few issues that need to be addressed. Unless courts and judiciaries address the still underexposed issues of impartiality, independence and governance, they may lose the legitimacy they so urgently need to deliver justice.